gao second generation antipsychotic drug use on poor kids

8/9/2019 GAO Second Generation Antipsychotic Drug Use on Poor Kids

1/40

Department of Health and Human Services

OFFICE OFINSPECTOR GENERAL

SECOND

-GENERATION

ANTIPSYCHOTIC DRUG USE

AMONG MEDICAID-ENROLLED

CHILDREN

: QUALITY-OF-CARE

CONCERNS

Daniel R. Levinson

Inspector General

March 2015

OEI-07-12-00320


2/40

EXECUTIVE SUMMARY: SECOND-GENERATION ANTIPSYCHOTIC DRUG USE AMONGMEDICAID-ENROLLED CHILDREN: QUALITY-OF-CARE CONCERNSOEI-07-12-00320

WHY WE DID THIS STUDY

Second-generation antipsychotics (SGAs) are a class of drugs used to treat psychiatric disorders, such as

schizophrenia, bipolar disorder, and psychotic depression. SGAs are widely used to treat children enrolledin Medicaid who have mental health conditions. However, SGAs can have serious side effects and littleclinical research has been conducted on the safety of treating children with these drugs. Consequently,children’s treatment with SGAs needs careful management and monitoring. This evaluation examines thequality of care provided to children receiving SGAs that were paid for by Medicaid.

HOW WE DID THIS STUDY

We selected a sample of 687 claims for SGAs prescribed to children in California, Florida, Illinois, New York, and Texas. These States represented approximately 39 percent of total Medicaid payments forSGAs in 2011. Board-certified child and adolescent psychiatrists reviewed medical records related to thesampled claims using seven criteria related to quality-of-care concerns (see the chart below for the criteria).

We established these criteria on the basis of information and guidelines issued by various Federal and Stateagencies and professional associations regarding the prescribing of psychotropic drugs to children.

WHAT WE FOUND

The graphic below illustrates the percentages of quality-of-care concerns identified in the medical recordsassociated with claims for SGAs prescribed to children.

Medical reviewers identified multiple quality-of-care concerns for many claims.

Source: OIG analysis of Medicaid claims in five states, 2014.


3/40

In the five States, 8 percent of SGAs were prescribed for the limited number of medically accepted pediatric indications. There are only five SGAs with medically accepted pediatric indications. It is notuncommon for doctors to prescribe, or Medicaid to pay for, SGAs for children for indications that are notmedically accepted. Medically accepted indications include both uses of drugs approved by the Food andDrug Administration (FDA) and uses supported by one or more of three drug compendia. It is difficult toconduct the clinical trials needed to obtain FDA approval or compendia support for pediatric uses of drugs.

Three of the eleven SGAs carry an FDA boxed warning regarding increased chances of suicidal thinkingand behavior in pediatric patients. We found that over a third of SGAs were prescribed in the presence ofconditions described in the FDA boxed warning. Physicians are not prohibited from prescribing a drug fora patient who has the condition(s) specified in the FDA boxed warning if the physician judges that the benefits may outweigh the risks.

WHAT WE RECOMMEND

To ensure the quality of the care provided to children receiving SGAs, we made three recommendations tothe Centers for Medicare & Medicaid Services (CMS). First, we recommended that CMS work with StateMedicaid programs to perform utilization reviews of SGAs prescribed to children. Second, we

recommended that CMS work with State Medicaid programs to conduct periodic reviews of medicalrecords associated with claims for SGAs prescribed to children. Third, we recommended that CMS workwith States to consider other methods of enhanced oversight of SGAs prescribed to children, such asimplementing peer review programs. CMS concurred with all three recommendations.


4/40

TABLE OF CONTENTS

Objectives ....................................................................................................1

Background..................................................................................................1

Methodology................................................................................................6

Findings........................................................................................................9

In the five States studied, quality-of-care concerns were

identified in the medical records for 67 percent of claims for

SGAs prescribed to children ............................................................9

In the five States studied, 8 percent of SGAs were prescribed

for the limited number of medically accepted pediatric uses ........12

Conclusion and Recommendations ............................................................15

Agency Comments and Office of Inspector General Response .....16

Appendixes ................................................................................................17

A: FDA Boxed Warning Regarding Children on SGA

Product Labeling............................................................................17

B: Sources of Guidance and Information on the Use of SGAs

To Treat Children ...........................................................................18

C: Detailed Methodology..............................................................21

D: Descriptive Data on Population and Sampled Claims .............26

E: Statistical Estimates and Confidence Intervals ........................29

F: Agency Comments ....................................................................31

Acknowledgments......................................................................................35


5/40

OBJECTIVES

To determine the extent to which claims for second-generation antipsychotics

(SGAs) prescribed to children enrolled in Medicaid in five selected States:

1.

presented quality-of-care concerns and

2. were prescribed for indications other than medically accepted indications

and/or in the presence of conditions specified in the Food and Drug

Administration (FDA) boxed warning.

BACKGROUND

FDA Approval of SGAs

FDA must approve drugs before companies can legally sell them in the United

States.1 If FDA determines that a drug’s benefits for its intended use outweigh its

known risks, then FDA approves the drug for sale.2 At the time of our review,FDA had approved five SGAs for use by children in the treatment of

schizophrenia, bipolar disorder, and irritability associated with autistic disorder.3, 4

The five SGAs that FDA has approved for use by children are aripiprazole,

olanzapine, paliperidone, quetiapine fumarate, and risperidone. Aripiprazole and

risperidone are approved to treat all three of the diagnoses named above.

Olanzapine and quetiapine fumarate are approved to treat schizophrenia and

bipolar disorder. Paliperidone is approved to treat only schizophrenia. There are

six additional SGAs with FDA-approved uses for adults.5

Off-label uses of FDA-approved drugs. After FDA approves a drug for a specific

use, physicians are permitted to prescribe that drug for other uses. Prescribingdrugs for nonapproved uses is often referred to as “off-label prescribing.”

Off-label prescribing is not uncommon, especially for drugs prescribed to

____________________________________________________________

1 FDA approval of a drug is based on the results of testing for safety and efficacy by the companywishing to sell the drug. FDA, Development and Approval Process (Drugs). Accessed athttp://www.fda.gov/Drugs/DevelopmentApprovalProcess/default.htm on May 12, 2014.2 FDA, Approved Drugs: Questions and Answers. Accessed athttp://www.fda.gov/drugs/resourcesforyou/consumers/ucm054420.htm on May 12, 2014.3

Antipsychotic drugs are separated into two classes: first-generation (or conventional) andsecond-generation (or atypical). First-generation antipsychotic drugs were introduced in the1950s; SGAs were introduced in the 1990s.4 These five SGAs are approved by FDA for use by children under age 18. For purposes of thisstudy, we defined “children” to include those 0–17 years old during the review period to conformto the FDA’s age ranges.5 The six additional SGAs with FDA-approved uses for adults are asenapine, clozapine,iloperidone, lurasidone, olanzapine/fluoxetine, and ziprasidone. FDA, Atypical Antipsychotic Drugs Information. Accessed at http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm094303.htm on May 12, 2014.

SGA Use Among Medicaid-Enrolled Children: Quality-of-Care Concerns (OEI-07-12-00320) 1

http://www.fda.gov/Drugs/DevelopmentApprovalProcess/default.htmhttp://www.fda.gov/drugs/resourcesforyou/consumers/ucm054420.htmhttp://www.fda.gov/Drugs/DrugSafetyhttp://www.fda.gov/Drugs/DevelopmentApprovalProcess/default.htmhttp://www.fda.gov/drugs/resourcesforyou/consumers/ucm054420.htmhttp://www.fda.gov/Drugs/DrugSafety


6/40

children, as it is difficult to conduct the clinical trials needed for FDA approval of

drugs for the pediatric population.6

FDA boxed warning. If drug manufacturers and/or FDA determine that a drug

may have severe or life-threatening risks when used to treat patients with specific

conditions, FDA requires that the product’s labeling include a boxed warning

describing those risks.7 However, physicians are not prohibited from prescribing

a drug for a patient who has the condition(s) specified in the FDA boxed warning

if the physician judges that the benefits may outweigh the risks.

As of June 2012, three SGAs—aripiprazole, olanzapine/fluoxetine, and quetiapine

fumarate—carried an FDA boxed warning specific to children. The warning

describes the increased chance of suicidal thinking and behavior in children,

adolescents, and young adults with major depressive disorder and other

psychiatric disorders being treated with them. Two of these SGAs that carry the

boxed warning—aripiprazole and quetiapine fumarate—are also among the five

SGAs that have FDA-approved pediatric uses. See Appendix A for an example ofthe boxed warning for an SGA.

Medicaid Coverage of Outpatient Prescription Drugs

All State Medicaid programs cover outpatient prescription drugs.8 State Medicaid

programs must pay for covered outpatient drugs prescribed for medically accepted

indications and may pay for outpatient drugs not prescribed for medically

accepted indications.

Medically accepted indications. The Act defines “medically accepted

indications” as both the uses approved by FDA and those uses, including off-label

uses, that are supported by one or more of three compendia.9 At the time of our

review, only one SGA—risperidone—had any medically accepted indications for

pediatric use beyond those approved by FDA.10 For the other SGAs, there were

____________________________________________________________

6 FDA, Drug Research and Children. Accessed at http://www.fda.gov/Drugs/ResourcesForYou/Consumers/ucm143565.htm on April 22, 2014. Institute of Medicine (US) Committee on ClinicalResearch Involving Children, Field and Berman, eds., Ethical Conduct of Clinical Research Involving Children, Washington, DC: National Academies Press, 2004; pp. 59 and 61. Accessedat http://www.ncbi.nlm.nih.gov/books/NBK25557/pdf/TOC.pdf on April 22, 2014.7 FDA, An FDA Guide to Drug Safety Terms. Accessed at

http://www.fda.gov/forconsumers/consumerupdates/ucm107970.htm on May 12, 2014.8 Social Security Act (the Act) § 1905(a)(12); CMS, Medicaid Prescription Drugs Program.Accessed at http://www.medicaid.gov/Medicaid-CHIP-Program-Information/By-Topics/Benefits/Prescription-Drugs/Prescription-Drugs.html on May 12, 2014.9 Section 1927(g)(1)(B)(i)(I)–(III) of the Act.

The three compendia are (1) the American Societyof Health System Pharmacists, Inc.’s American Hospital Formulary Service Drug Information,(2) the United States Pharmacopeia—Drug Information (or its successor publications), and(3) DrugDEX Information System.10 Medically accepted indications for risperidone include its FDA-approved uses as well as its useto treat (1) behavioral syndrome-mental retardation and (2) pervasive developmental disorder.


http://www.fda.gov/Drugs/ResourcesForYouhttp://www.ncbi.nlm.nih.gov/books/NBK25557/pdf/TOC.pdfhttp://www.fda.gov/forconsumers/consumerupdates/ucm107970.htmhttp://www.medicaid.gov/Medicaid-CHIP-Program-Information/Byhttp://www.fda.gov/Drugs/ResourcesForYouhttp://www.ncbi.nlm.nih.gov/books/NBK25557/pdf/TOC.pdfhttp://www.fda.gov/forconsumers/consumerupdates/ucm107970.htmhttp://www.medicaid.gov/Medicaid-CHIP-Program-Information/By


7/40

no medically accepted indications beyond the FDA-approved uses; the medically

accepted indications and the FDA-approved uses are identical.

Limitations on Medicaid coverage of drugs. Section 1927 of the Act establishes

the permissible limitations on and exclusions of Medicaid coverage of outpatient

drugs.11 Under section 1927, States can choose to limit or exclude coverage of

outpatient drugs if the prescribed uses are not for medically accepted indications.12

However, State Medicaid programs may choose to cover outpatient drugs not

prescribed for medically accepted indications. Table 1 shows which SGAs have

FDA-approved pediatric uses, which SGAs have medically accepted pediatric

indications other than the FDA-approved pediatric uses, and which SGAs carry

the FDA boxed warning specific to children.

Table 1: FDA-Approved Pediatri c Uses, Medically Accepted Pediatric Indications,and FDA Boxed Warnings on Second-Generation Antipsychotics

Drug NameFDA-Approved

Pediatric Use(s)

Medically Accepted Pediatric

Indication(s) Other ThanFDA-Approved Pediatric Use(s)

FDA Boxed WarningSpecific to Children

aripiprazole

asenapine

clozapine

iloperidone

lurasidone

olanzapine

olanzapine/fluoxetine

paliperidone

quetiapine fumarate

risperidone

ziprasidone

Yes

No

No

No

No

Yes

No

Yes

Yes

Yes

No

No

No

No

No

No

No

No

No

No

Yes

No

Yes

No

No

No

No

No

Yes

No

Yes

No

No

Source: Office of Inspector General (OIG) analysis of compendia, 2014.

Four of the five States included in this evaluation—California, Florida, Illinois,

and Texas—covered drugs not prescribed for medically accepted indications.

These States used information regarding medically accepted indications in

making case-by-case coverage decisions, such as decisions on requests for prior

____________________________________________________________

11 Section 1927(k)(2) of the Act defines “covered outpatient drugs” as drugs that are dispensedonly upon prescription and that are approved for safety and effectiveness as prescription drugsunder the Federal Food, Drug, and Cosmetic Act.12 Section 1927(d)(1)(B)(i) of the Act.


http:///reader/full/drugs.11http:///reader/full/drugs.11http:///reader/full/indications.12http:///reader/full/indications.12http:///reader/full/drugs.11http:///reader/full/indications.12


8/40

authorization.13 Of the five States reviewed, only New York limited coverage of

drugs to those prescribed for medically accepted indications.

Sources of Guidance and Information on the Use of SGAs To Treat

Children

There are no Federal Medicaid requirements regarding the appropriate prescribing

of SGAs to children. However, a number of Federal and State agencies and

professional associations have issued information and guidelines on the

prescribing of such drugs to children.14 These guidelines and information include:

AACAP practice parameter —The American Academy of Child and

Adolescent Psychiatry’s (AACAP) practice parameter on the use of

psychotropic medications to treat children identifies practices that may

reduce the use of ineffective and inappropriate medications. These

practices include developing monitoring plans, using adequate dosages

and durations of treatment, and providing a clear rationale for the use of

polypharmacy (the concurrent use of multiple drugs).15

HEDIS measures —The National Committee for Quality Assurance

(NCQA) publishes the Healthcare Effectiveness Data and Information Set

(HEDIS), which is used to measure the performance of health plans.16 In

July 2014, NCQA published new measures for HEDIS 2015 on the safe

and judicious use of antipsychotics in children and adolescents. These

measures assess the appropriate use and management of antipsychotic

drugs in youth. Reporting on the new measures will begin in June 2015.

ACF information memorandum —On April 11, 2012, the Administration

for Children and Families (ACF) released an information memorandum

entitled “Promoting the Safe, Appropriate, and Effective Use of

Psychotropic Medication for Children in Foster Care.”17 This

memorandum identifies three “outlier practices” that may indicate that

issues other than clinical need are influencing the prescribing of SGAs:

____________________________________________________________

13 States may require, as a condition of coverage or payment for a covered outpatient drug, thata drug be authorized before it is dispensed. This type of approval is known as prior authorization.Sections 1927(d)(1)(A) and (d)(5) of the Act.14

The guidance discussed here covers all psychotropic drugs—a category that includes SGAs andseveral other types of drugs. Other classes of psychotropic drugs include attention deficit/hyperactivity disorder (ADHD) drugs, antianxiety drugs, antidepressants, first-generationantipsychotics, hypnotics, and mood stabilizers. This study focuses specifically on SGAs.15 AACAP, “Practice Parameter on the Use of Psychotropic Medication in Children andAdolescents,” J. Am. Acad. Child Adolesc. Psychiatry, 2009; 48(9): 961–973.16 State Medicaid programs may report voluntarily on the HEDIS measures.17 ACF, Promoting the Safe, Appropriate, and Effective Use of Psychotropic Medication forChildren in Foster Care, Log No. ACYF-CB-IM-12-03, April 11, 2012. Accessed athttp://www.acf.hhs.gov/sites/default/files/cb/im1203.pdf on May 12, 2014.


http:///reader/full/authorization.13http:///reader/full/authorization.13http:///reader/full/children.14http:///reader/full/children.14http:///reader/full/drugs).15http:///reader/full/drugs).15http:///reader/full/plans.16http:///reader/full/plans.16http://www.acf.hhs.gov/sites/default/files/cb/im1203.pdfhttp:///reader/full/authorization.13http:///reader/full/children.14http:///reader/full/drugs).15http:///reader/full/plans.16http://www.acf.hhs.gov/sites/default/files/cb/im1203.pdf


9/40

prescribing too many concurrent drugs, prescribing dosages that are too

high, and prescribing SGAs to children at too young an age.

Report and Resource Guide —In 2010, a collaboration among 16 States,

the Rutgers University Center for Education and Research on Mental

Health Therapeutics (CERTs), and the Medicaid Medical Directors

Learning Network (MMDLN) produced the Antipsychotic Medication Use

in Medicaid Children and Adolescents: Report and Resource Guide From

a 16-State Study (hereinafter referred to as the Report and Resource

Guide).18 This guide contains information on utilization patterns and

treatment practices that raise clinical concerns. Among the issues listed

are polypharmacy, wrong dosages, and the prescribing of antipsychotic

drugs to very young children.19

State utilization guidelines —State agencies in Florida and Texas have

each developed utilization guidelines that can help identify when further

review of the treatment of children with SGAs is necessary. These States’guidelines address issues such as polypharmacy, inadequate indications for

use of drugs, excessive dosage, and prescribing of drugs to very young

children.

For detailed information on these sources of guidance, see Appendix B.

Related Studies

A 2011 Government Accountability Office (GAO) report found that in 2008,

children in foster care in five selected States were prescribed psychotropic drugs,

including antipsychotic drugs, at higher rates than children not in foster care .20

The GAO report states that the higher rates of usage do not necessarily indicate

inappropriate prescribing practices and notes that these higher rates could be due

to the greater mental health needs of children in foster care and the challenges of

coordinating their medical care. According to the GAO report, hundreds of

children in the selected States had drug regimens consisting of five or more

psychotropic drugs, although there is no medical evidence to support the

concurrent use of this many drugs in children.

____________________________________________________________

18 CERTs and MMDLN are both funded by the Agency for Healthcare Research and Quality. The

16 States were Alabama, California, Colorado, Illinois, Indiana, Maine, Massachusetts, Missouri, New Hampshire, New York, Oklahoma, Oregon, Pennsylvania, Tennessee, Texas, andWashington. Of these, California, Illinois, New York, and Texas are among the five Statesincluded in this evaluation.19 Antipsychotic Medication Use in Medicaid Children and Adolescents: Report and ResourceGuide From a 16-State Study. MMDLN/Rutgers CERTs Publication No. 1, June 2010.20 The five States included in GAO’s study were Florida, Massachusetts, Michigan, Oregon, andTexas. The children studied—both those in foster care and not in foster care—were enrolled inthe Medicaid programs of the five States. GAO, Foster Children: HHS Guidance Could HelpStates Improve Oversight of Psychotropic Prescriptions (GAO-12-201), December 2011.


http:///reader/full/Guide).18http:///reader/full/Guide).18http:///reader/full/Guide).18http:///reader/full/children.19http:///reader/full/children.19http:///reader/full/Guide).18http:///reader/full/children.19


10/40

In 2011, OIG released a study on elderly nursing home residents’ use of SGAs.

This study determined that 14 percent of elderly nursing home residents had

Medicare claims for SGAs.21 Eighty-three percent of Medicare claims for SGAs

for elderly nursing home residents were associated with off-label conditions, and

88 percent were associated with the condition specified in the FDA boxed

warning (i.e., dementia).

METHODOLOGY

State Selection

We purposively selected five States for inclusion in this evaluation: California,

Florida, Illinois, New York, and Texas. These States were selected because they

represented approximately 39 percent of total Medicaid payments for SGAs in

2011.22

From these five States, we collected all fee-for-service Medicaid-paid claims for

SGAs prescribed to children aged 0–17 years old between January 1 andJune 30, 2011.23 We also collected information from each State regarding which

children were enrolled in foster care and information on the States’ drug coverage

policies.

To reduce the overall sample size, answer specific questions, and achieve precise

estimates, we stratified the population into four groups based on foster care

enrollment and payment amount. We selected a stratified random sample of

687 claims. For a presentation of the stratified sample and detailed

methodological information, see Appendix C. For detailed descriptive

information on the population of claims and the sampled claims, see Appendix D.

Development of Criteria for Medical Record Review. Because there are no

Federal Medicaid requirements defining appropriate use of SGAs to treat

children, we reviewed existing guidance and literature to identify criteria for

determining whether quality-of-care concerns were present in children’s treatment

with SGAs. (Specifically, we reviewed all of the sources listed on pages 4 and 5

except for the HEDIS 2015 measures, which at the time had not yet been

published.) We identified seven criteria related to quality of care in the treatment

of children with SGAs; each criterion is supported by at least three of the sources

____________________________________________________________

21 OIG, Medicare Atypical Antipsychotic Drug Claims for Elderly Nursing Home Residents(OEI-07-08-00150), May 2011.22 The 39-percent figure is based on 2011 claims data from the Medicaid Statistical Information

System (MSIS).

23 The claims data collected included only fee-for-service claims. Initial discussions withMedicaid staff indicated that three of the five selected States had no managed-care coverage of SGAs for children during our review period. The other two States indicated that managed-carecoverage of these drugs for children represented a very small percentage of paid claims for drugsin the review period.



11/40

of guidance previously described. Table 2 describes each of the seven

quality-of-care criteria we identified and shows which of the sources of guidance

and information that we reviewed address the issue.24

Table 2: Descripti ons of and Support for Medical Record Review Criteria and CorrespondingQuality-of-Care Concerns

Criterion

Source

AACAP

Practice

Parameter

ACF

Memo

Report and

Resource

Guide

Florida Utilization

Guidelines

Texas Utilization

Guidelines

Dosage (wrong dose)

Duration (taken too long)

Indications for use (wrong treatment)

Monitoring (poor monitoring)

Polypharmacy (too many drugs)

Patient age (too young)

Side effects

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

ource: OIG analysis of guidance documents, 2014.

Review of Medical Records and Claims Data

The psychiatrists who performed the medical reviews were board-certified child

and adolescent psychiatrists with extensive experience and expertise in treating

children with mental health conditions. To ensure a consistent interpretation of

the seven criteria, we provided the contracted psychiatrists with the same sourcesof guidance and information that we used to develop the criteria. The contracted

psychiatrists determined whether quality-of-care concerns existed in each child’s

SGA treatment, assessing each of the seven criteria using their clinical experience

and judgment. They also identified the diagnosis/diagnoses or indication(s) that

the SGAs were prescribed to treat. OIG staff used this diagnosis information, as

well the children’s ages, to determine whether the drugs were prescribed for

medically accepted indications.

We received records for 485 of the 687 sampled claims for an overall weighted

response rate of 75 percent. The remaining 202 claims were undocumented or

insufficiently documented, or the providers could not be located to request the

records. Specifically, we could not locate or contact the providers or facilities

that had custody of the medical records for 170 sampled claims. Although we

____________________________________________________________

24 Because the new HEDIS measures regarding the use of antipsychotic drugs by children had not been published at the time of our review, we could not include them in the sources of guidance weused to develop the medical record review criteria.


http:///reader/full/issue.24http:///reader/full/issue.24http:///reader/full/issue.24


12/40

were able to contact the providers, we did not receive any medical records for

7 sampled claims, and we received insufficient medical records for an additional

25 claims. See Appendix C for details on the response rates and a nonresponse

analysis.

Limitations

It is possible that the children in our sample received SGAs that were not paid for

by Medicaid; therefore, these children’s use of SGAs may be underestimated.

Additionally, the psychiatrists’ reviews were based solely on the medical record

documentation received; some of the quality-of-care concerns identified could be

the result of poor documentation rather than poor quality of care.

Standards

This study was conducted in accordance with the Quality Standards for Inspection

and Evaluation issued by the Council of the Inspectors General on Integrity and

Efficiency.



13/40

FINDINGS

In the five States studied, quality-of-care concerns wereidentified in the medical records for 67 percent of claims forSGAs prescribed to children

Quality-of-care concerns were identified in the medical records associated with

67 percent of claims for SGAs prescribed to children; two or more quality-of-care

concerns were identified in the medical records for 49 percent of claims. Chart 1

below illustrates the percentages of claims for which concerns related to each

medical record review criterion were identified in medical records; see Table D-5

in Appendix D for estimated totals and percentages of claims with each

quality-of-care concern. See Appendix E for all estimates and 95-percent

confidence intervals for projections.

Chart 1: Percentages of Medicaid Claims With Quality-of-Care ConcernsRegarding Children’s Use of SGAs in 2011

Medical reviewers identified multiple quality-of-care concerns for many claims.

Source: OIG analysis of Medicaid claims in five States, 2014.



14/40

Medical reviewers identified many specific concerns related to each of the seven

criteria for the medical record review. Below are four case studies that illustrate

some of the quality-of-care concerns that medical reviewers identified regarding

each criterion.

Case Study 1 – a 10-year-old child diagnosed with ADHD

Monitoring —The medical record lacked documentation of monitoring for side

effects, including a lack of recognition of or monitoring for akathisia.25

Indications for use (wrong treatment) —This child’s ADHD was clearly

documented as being treated with a stimulant medication, but there was no

documented reason for the child’s treatment with the SGA.

Polypharmacy (too many drugs) —This child was prescribed three psychotropic

drugs during the review period, one of which was an SGA.

Duration (taken too long) —The medical reviewer stated: “The [SGA] should

have been tapered down and possibly discontinued completely.”

Dosage (wrong dose) —This child presented symptoms that appeared to worsen

with the prescribed course of treatment with the SGA. When the dosage of the

SGA was increased, the child’s symptoms worsened. The pattern of increasing

dosage and worsening symptoms was repeated several times throughout the

review period.

Side effects —The medical reviewer stated: “It is very possible the [SGA] caused

the side effect of akathisia, and increasing the dosage caused the patient to get

worse instead of better.”

Case Study 2 – a 4-year-old child diagnosed with ADHD and a mood disorder

Monitoring —There was no evidence in the child’s medical records of any

monitoring while the child was taking the sampled SGA.

Indications for use (wrong treatment) —This child was placed in the custody of

child protective services at 4 weeks old and had a history of abuse and neglect and

severe psychosocial issues, such as aggression, anxiety, and mood swings. The

reviewer stated that individual, family, and behavior therapy should have been

attempted before initiating treatment with drugs. However, there was no evidence

in the child’s medical record indicating that such therapies were attempted.

Polypharmacy (too many drugs) —This child was prescribed four psychotropicdrugs during the review period, of which two were SGAs.

____________________________________________________________

25 Akathisia is “a pathologic condition characterized by restlessness and agitation, such as aninability to sit still,” and is a common side effect of many SGAs. Mosby’s Medical Dictionary,7th Ed. Mosby-Elsevier, 2006.


http:///reader/full/akathisia.25http:///reader/full/akathisia.25http:///reader/full/akathisia.25


15/40

Dosage (wrong dose) —The reviewer noted that there is no appropriate dosage of

the prescribed SGA for this child’s conditions.

Age (too young) —The reviewer stated that treatment with SGAs was not

appropriate for a 4-year-old.

Case Study 3 – a 16-year-old child diagnosed with bipolar disorder

Monitoring —This child experienced significant weight gain while taking the

SGA, but there was no evidence in the child’s record of monitoring of blood

sugar, cholesterol, or other hormone levels in response to the weight gain.

Indications for use (wrong treatment) —The reviewer stated that the diagnosis

which the SGA was prescribed to treat was poorly documented, saying, “Multiple

antipsychotic agents were used with vague documentation of hallucinations [as

the only explanation].”

Polypharmacy (too many drugs) —This child was prescribed a total of six

psychotropic drugs during the review period, of which three were SGAs.

Duration (taken too long)—The SGA was prescribed to this child in response to

a report of hallucinations. However, the medical reviewer stated: “When all

medications were discontinued… hallucinations were not reported nor was the

patient reporting suicidal ideation [i.e., suicidal thinking].” The SGA should have

been discontinued as soon as it became apparent that its side effects were

outweighing any treatment benefit.

Side effects —This child experienced paranoia, hostility, unstable mood,

hallucinations, and suicidal thoughts. This child also experienced significant side

effects potentially resulting from the prescribed drugs, including a 22-poundweight gain, insomnia, and edema (swelling) of hands and feet.

Case Study 4 – a 16-year-old child diagnosed with bipolar disorder and ADHD

Monitoring —Although the child’s medical record indicated that laboratory

testing was performed, there were no notes in the record of the results of that

testing. The medical record also lacked any evidence of monitoring of the child’s

height, weight, or vital signs.

Polypharmacy (too many drugs) —This child was prescribed a total of six

psychotropic drugs during the review period, of which two were SGAs. The

medical reviewer noted that although the regimen of six psychotropic drugs mighthave been beneficial in treating the child’s diagnoses, the medical record lacked

documentation to support why any of the drugs were chosen to treat the child and

the symptoms these drugs were targeting.

Dosage (wrong dose) —This child was prescribed double the maximum

recommended dosage of the SGA to treat bipolar disorder.



16/40

Lack of monitoring was the most commonly identified quality-of-care

issue

In the five States we reviewed, quality-of-care concerns regarding lack of

monitoring were identified in the medical records for 53 percent of claims for

SGAs prescribed to children. Reviewers were concerned about a lack of

monitoring for many physiological and behavioral changes. They especially

noted that the following were not being regularly performed: measuring of

height, weight, vital signs, and blood pressure; measuring of abnormal

involuntary movements; laboratory testing (e.g., tests of liver function, measures

of blood glucose levels, measures of lipid levels); and electrocardiograms. In all

four case studies described above, reviewers identified a lack of monitoring,

including a failure to recognize or manage side effects, such as akathisia,

significant weight gain, insomnia, and edema.

In the five States studied, 8 percent of SGAs were

prescribed for the limited number of medically acceptedpediatric indications

Eight percent of claims for SGAs were for drugs prescribed for the limited

number of medically accepted pediatric indications. As noted in the background,

there are only five SGAs with medically accepted pediatric indications; most of

those indications are limited to specific age ranges for each drug. Furthermore,

physicians are not prohibited from prescribing SGAs for indications that are not

medically accepted. See Table 3 for a list of medically accepted pediatric

indications of SGAs.

Among the 92 percent of claims that were not prescribed for medically accepted

pediatric indications, common diagnoses that SGAs were prescribed to treat were

bipolar disorder (20 percent), mood disorders (13 percent), and autism spectrum

disorders (8 percent).26

____________________________________________________________

26 Although there are medically accepted indications for certain drugs to treat bipolar disorder andautistic disorder, the children who were prescribed these drugs for nonmedically acceptedindications were (1) not in the age range specified and/or (2) were not diagnosed with thesymptoms specified. For example, a child diagnosed with bipolar disorder may not have hadmania or mixed episodes; a child diagnosed with autistic disorder may not have had irritability.


http:///reader/full/percent).26http:///reader/full/percent).26http:///reader/full/percent).26


17/40

Table 3: Medically Accepted Pediatric Indications for SGAs

SGA Medically Accepted Pediatric Indication(s)1 Age Range

aripiprazole

olanzapine

paliperidone

quetiapine fumarate

risperidone

Schizophrenia

Mania or mixed episodes associated with bipolar Idisorder 2

Irritability associated with autistic disorder

Schizophrenia

13–17 years

10–17 years

6–17 years

13–17 years

Mania or mixed episodes associated with bipolar I disorder

Schizophrenia

Schizophrenia

Mania associated with bipolar I disorder

Schizophrenia

Mania or mixed episodes associated with bipolar I disorder

Irritability associated with autistic disorder

Behavioral syndrome-mental retardation

Pervasive developmental disorder

13–17 years

12–17 years

13–17 years

10–17 years

13–17 years

10–17 years

5–16 years

--3

--3

1Medically accepted indications include both the uses approved by FDA and those uses, including off-label uses,supported by one or more of the compendia.2 Bipolar disorder is a brain disorder that causes unusual shifts in mood, energy, activity levels, and the ability toperform daily tasks. There are two primary forms of bipolar disorder: bipolar I disorder and bipolar II disorder. Eachof these two types of bipolar disorder have distinct and separate symptoms.3The FDA-approved uses for risperidone have specific age ranges; the compendia-supported indications do not havespecific age ranges, but are simply noted as “pediatric use.”

Source: OIG analysis of compendia information, 2014.

New York had a coverage policy stating that its Medicaid program would pay

only for SGAs prescribed for medically accepted indications. However, at least3,366 claims totaling $773,607 were paid in violation of this policy.27 New York

State Medicaid staff told us that because of the lack of diagnosis information on

drug claims, this coverage policy is difficult to enforce without conducting a

medical record review.

Over a third of SGAs were prescribed for condit ions described in the

FDA boxed warning

For drugs that may have severe or life-threatening risks when used to treat

patients with specific conditions, FDA may require that the product’s labeling

____________________________________________________________27 To determine whether a drug is prescribed for a medically accepted pediatric indication, wemust know the diagnosis it was prescribed to treat. However, diagnosis information is notincluded in drug claims data; it is available only from medical records. To identify claims that didnot meet New York’s coverage policy, we identified all the claims in New York’s population datathat were for the five SGAs that have no medically accepted pediatric indications: aripiprazole, paliperidone, risperidone, quetiapine fumarate, and olanzapine. There could be additional claimsin the population data that were not prescribed for medically accepted pediatric indications andthus violate New York’s coverage policy but that we cannot identify without diagnosisinformation and/or medical review.


http:///reader/full/policy.27http:///reader/full/policy.27http:///reader/full/policy.27


18/40

include a boxed warning describing those risks. Three of the SGAs included in

this evaluation carried an FDA boxed warning describing the increased chance of

suicidal thinking and behavior in children, adolescents, and young adults with

major depressive disorder and other psychiatric disorders being treated with these

SGAs. However, physicians are not prohibited from prescribing a drug for a

patient who has the condition(s) specified in the FDA boxed warning if the physician judges that the benefits may outweigh the risks.

Thirty-seven percent of claims for SGAs were prescribed for conditions described

in the FDA boxed warning—i.e., for major depressive disorders and other

psychiatric disorders. The language of the warning—particularly the phrase

“other psychiatric disorders”—is broad and could apply to almost any child with a

mental health condition. All children taking the three SGAs that carry this

warning, regardless of their exact diagnoses or conditions, should be appropriately

monitored for suicidal thoughts and other side effects.



19/40

CONCLUSION AND RECOMMENDATIONS

SGAs are widely used to treat children enrolled in Medicaid who have mental

health conditions. However, SGAs have been previously found to cause serious

side effects and little clinical research has been conducted on the treatment of

children with these drugs. Quality-of-care concerns were identified in the medicalrecords for 67 percent of claims for SGAs prescribed to children. Lack of

monitoring was the most commonly identified quality-of-care issue, being

identified in just over half of the claims for SGAs.

The high percentage of claims for which our reviewers identified quality-of-care

concerns indicates that more needs to be done to ensure the quality of care

provided to children receiving SGAs paid for by Medicaid. To ensure the quality

of the care provided to these children, we recommend that CMS work with State

Medicaid programs to:

Perform utili zation reviews of SGAs prescribed to children

Utilization reviews could specifically focus on the children’s age, the duration of

their treatment with SGAs, and their overall drug regimens. The previously

described guidance and information on using SGAs to treat children—such as the

utilization guidelines developed by the Florida and Texas Medicaid programs—

may be of use to CMS and to other State Medicaid programs in developing

guidelines for such utilization reviews. The medical review criteria that OIG

developed for this evaluation may also be helpful in developing such guidelines.

Conduct periodic reviews of medical records associated with claims

for SGAs prescribed to chi ldren

Periodic reviews of medical records could ensure that (1) there are clear rationales

for prescribing the SGAs, (2) children are being properly monitored, and

(3) children’s dosages are properly adjusted.

Consider other methods of enhanced oversight of SGAs prescribed

to children

Some of the other methods of enhanced oversight that CMS and State Medicaid

programs could consider include:

implementing peer review programs through which prescribers encourage

one another to improve quality of care and

undertaking voluntary reporting of the HEDIS measures regarding

children’s use of antipsychotic drugs or adopting the HEDIS measures in

State oversight of SGAs.

In addition to making these recommendations, we have forwarded to CMS the

information on the undocumented claims identified in our sample, so that it may

take appropriate action.



20/40

AGENCY COMMENTS AND OFFICE OF INSPECTOR GENERAL

RESPONSE

CMS concurred with all three of our recommendations.

In responding to our first recommendation, CMS stated that it would work with

States through the Medicaid Drug Utilization Review Program to monitorchildren’s use of antipsychotic drugs and provide feedback to States regarding any

quality-of-care concerns identified using the criteria we established.

In responding to our second recommendation, CMS stated that it would encourage

States with managed care programs to request that their External Quality Review

Organizations conduct periodic reviews of medical records of children treated

with SGAs. We suggest that CMS also consider methods of periodically

reviewing medical records of children who receive SGAs paid through

fee-for-service arrangements.

In response to our third recommendation, CMS plans to take a number of actions.In general, CMS will:

share our report to spur conversation about how to promote safe treatment

of children with SGAs,

promote best practices in appropriate oversight of treatment of children

with mental or behavioral health issues,

use measures regarding the safe and judicious treatment of children and

adolescents with antipsychotics, and

engage with other Federal agencies to improve access to screening and

treatment for children with substance use or mental health disorders.

We support CMS’s effort to ensure quality of care in children’s treatment with

SGAs and its commitment to engaging with Federal partners and other

stakeholder organizations. For the full text of CMS’s comments, see Appendix F.



21/40

Antidepressants increased the risk compared to placebo of suicidal thinking and

behavior (suicidality) in children, adolescents, and young adults in short-term

studies of major depressive disorder (MDD) and other psychiatric disorders.

Anyone considering the use of [drug name] or any other antidepressant in achild, adolescent, or young adult must balance this risk with the clinical need.

Short-term studies did not show an increase in the risk of suicidality with

antidepressants compared to placebo in adults beyond age 24; there was a

reduction in risk with antidepressants compared to placebo in adults aged 65 and

older. Depression and certain other psychiatric disorders are themselves associated

with increases in the risk of suicide. Patients of all ages who are started on

antidepressant therapy should be monitored appropriately and observed closely

for clinical worsening, suicidality, or unusual changes in behavior. Families and

caregivers should be advised of the need for close observation and communication

with the prescriber. [Drug name] is not approved for use in pediatric patients with

depression [see WARNINGS AND PRECAUTIONS (5.2)].

APPENDIX A

FDA Boxed Warning Regarding Children on SGA Product Labeling

Three SGAs carry an FDA boxed warning regarding suicidality when used bychildren, adolescents, and young adults who have major depressive disorders or

other psychiatric disorders. In recent years, FDA has approved these three SGAs

to treat depression; therefore, the boxed warning has been applied to these SGAs

and refers to their use as antidepressants. The text of the warning, which is shown

in Figure A-1, is the same for all three SGAs.28

Figure A-1: SGA Boxed Warning Regarding Children

____________________________________________________________

28 The name of the drug has been elided from the warning.



22/40

APPENDIX B

Sources of Guidance and Information on the Use of SGAs To TreatChildren

To develop the criteria for the medical record review in this evaluation, we

reviewed the following documents issued by Federal and State agencies and by professional psychiatric associations.29

1. AACAP practice parameter. AACAP’s practice parameter identifies

practices that may reduce the use of ineffective and inappropriate

medications. These practices include conducting complete psychiatric and

medical evaluations, developing treatment and monitoring plans,

educating patients and their families about treatment and obtaining

consent for treatment, using adequate dosages and durations of treatment,

assessing responses to treatment, and providing a clear rationale for the

use of polypharmacy.

2. ACF information memorandum. On April 11, 2012, ACF released an

information memorandum titled “Promoting the Safe, Appropriate, and

Effective Use of Psychotropic Medication for Children in Foster Care.”30

This memorandum identifies three “outlier practices” that may indicate

that issues other than clinical need are influencing the prescribing of

SGAs, including:

Polypharmacy. ACF identifies the following patterns as being of

issue: use of three or more concurrent drugs, two or more concurrent

drugs in the same class for more than 30 days, and use of multiple

concurrent drugs before testing the effectiveness of a single drug.

Dosages of drugs that are too high. ACF notes that because so few

SGAs have received FDA approval for pediatric populations, there are

no recommended dosages for children for most SGAs. They note that

the lack of this information reinforces the need for effective

monitoring of children receiving SGAs for off-label uses.

Children being prescribed SGAs at too young an age. ACF notes the

dramatic rise in the use of SGAs among very young children and

concludes that because this population may be especially vulnerable to

adverse effects, careful management and oversight are necessary.

____________________________________________________________

29 The guidance discussed here covers all psychotropic drugs—a category that includes SGAs andseveral other types of drugs. Other classes of psychotropic drugs include ADHD drugs,antianxiety drugs, antidepressants, first-generation antipsychotics, hypnotics, and mood stabilizers.This study focuses specifically on SGAs.30 ACF, Promoting the Safe, Appropriate, and Effective Use of Psychotropic Medication forChildren in Foster Care, Log No. ACYF-CB-IM-12-03, April 11, 2012. Accessed athttp://www.acf.hhs.gov/sites/default/files/cb/im1203.pdf on May 12, 2014.


http:///reader/full/associations.29http:///reader/full/associations.29http://www.acf.hhs.gov/sites/default/files/cb/im1203.pdfhttp:///reader/full/associations.29http://www.acf.hhs.gov/sites/default/files/cb/im1203.pdf


23/40

3. Report and Resource Guide. The Rutgers University CERTs, the

MMDLN, and 16 States collaborated in 2010 to produce a publication for

State Medicaid programs and other stakeholders concerned with care of

children and youth. This publication is entitled Antipsychotic Medication

Use in Medicaid Children and Adolescents: Report and Resource Guide

From a 16-State Study.31 It contains background on issues about the use ofantipsychotic drugs by children, data from the 16 States reflecting use of

antipsychotic drugs by children, and information on how the 16 States are

managing the use of antipsychotic drugs.32 Among the issues listed are

polypharmacy, dosages, and the prescribing of antipsychotic drugs to very

young children.

4. State utilization guidelines. State agencies in Florida and Texas have

each developed utilization guidelines that can help identify when further

review of the prescribing of SGAs to children is necessary. Florida and

Texas’ guidelines address concerns such as polypharmacy (the concurrent

use of multiple drugs), inadequate indications for use of drugs, excessive

dosage, prescribing of drugs to very young children, and prescribing by a

primary care provider who has not received previous specialty training.33

NCQA publishes the HEDIS, a set of measures used to assess the performance

of health plans.34 In July 2014, NCQA adopted the new measures for HEDIS

2015 regarding the safe and judicious use of antipsychotics in children and

adolescents; reporting on the new HEDIS measures will begin in June 2015.

We did not use the HEDIS measures in developing our medical record review

criteria because the measures had not yet been adopted. We include

information on the measures here to provide additional context to our

____________________________________________________________

31 CERTs and MMDLN are both funded by the Agency for Healthcare Research and Quality.The 16 States were Alabama, California, Colorado, Illinois, Indiana, Massachusetts, Maine,Missouri, New Hampshire, New York, Oregon, Oklahoma, Pennsylvania, Tennessee, Texas, andWashington. Of these, California, Illinois, New York, and Texas are among the five States weincluded in this evaluation.32 Antipsychotic Medication Use in Medicaid Children and Adolescents: Report and ResourceGuide From a 16-State Study. MMDLN/Rutgers CERTs Publication No. 1, June 2010.33 University of South Florida, Bipolar I Acute or Mixed Mania—6–17 Years. Accessed athttp://medicaidmentalhealth.org/ViewGuideline.cfm?GuidelineID=32 on June 6, 2014. University

of South Florida, Principles of Practice Regarding the Use of Psychotropic Medication under Age 6 . Accessed at http://medicaidmentalhealth.org/ViewGuideline.cfm?GuidelineID=32 onMay 23, 2014. Texas Department of Family and Protective Services, Psychotropic MedicationUtilization Parameters for Foster Children, December 2010. Accessed athttp://www.dfps.state.tx.us/documents/about/pdf/txfostercareparameters-december2010.pdf onMay 12, 2014. Health Information Designs, Texas Prior Authorization Program—Clinical EditCriteria: Antipsychotics. Accessed at https://paxpress.txpa.hidinc.com/antipsychotics.pdf onMay 6, 2014.34 State Medicaid programs also report on the HEDIS measures, but their HEDIS reporting isvoluntary.


http:///reader/full/Study.31http:///reader/full/Study.31http:///reader/full/Study.31http:///reader/full/drugs.32http:///reader/full/drugs.32http:///reader/full/training.33http:///reader/full/training.33http:///reader/full/plans.34http:///reader/full/plans.34http://medicaidmentalhealth.org/ViewGuideline.cfm?GuidelineID=32http://medicaidmentalhealth.org/ViewGuideline.cfm?GuidelineID=32http://www.dfps.state.tx.us/documents/about/pdf/txfostercareparameters-december2010.pdfhttps://paxpress.txpa.hidinc.com/antipsychotics.pdfhttp:///reader/full/Study.31http:///reader/full/drugs.32http:///reader/full/training.33http:///reader/full/plans.34http://medicaidmentalhealth.org/ViewGuideline.cfm?GuidelineID=32http://medicaidmentalhealth.org/ViewGuideline.cfm?GuidelineID=32http://www.dfps.state.tx.us/documents/about/pdf/txfostercareparameters-december2010.pdfhttps://paxpress.txpa.hidinc.com/antipsychotics.pdf


24/40

recommendations. The HEDIS measures regarding the safe and judicious use

of antipsychotics in children and adolescents are:

use of higher-than-recommended doses of antipsychotics in children

and adolescents,

use of multiple concurrent antipsychotics in children and adolescents,

use of first-line psychosocial care for children and adolescents on

antipsychotics,

followup visits for children and adolescents on antipsychotics,

metabolic screenings for children and adolescents newly on

antipsychotics, and

metabolic monitoring for children and adolescents on antipsychotics.



25/40

APPENDIX C

Detailed Methodology

We stratified the sample by foster care enrollment, with the intention of

determining whether there were differences in quality of care between children

enrolled in foster care and those not enrolled in foster care. However, we wereunable to test for such differences because of the low response rate among the

providers associated with the claims in the foster care stratum, as described in the

section on response rates below. We also stratified the sample by payment

amount, in order to enhance the precision of our statistical estimates. Table C-1

provides details on the population and sample sizes by strata.

Table C-1: Population, Strata, and Sample Size

Stratum Claims in Population Claims in Sample

Claims for children who were in foster care on the service date ofthe claim

Claims for children who were not in foster care on the service dateof the claim and for which the paid amount was less than or equalto $187

Claims for children who were not in foster care on the service dateof the claim and for which the paid amount was greater than $187but did not exceed $557

Claims for children who were not in foster care on the service dateof the claim and for which the paid amount was greater than $557

49,058

173,596

99,419

28,460

147

92

267

181

Total 350,533 687

Source: OIG analysis of 2011 MMIS claims data from five States, 2014.

Identif ication of FDA-Approved Uses and Medically Accepted

Indications

To identify FDA-approved uses and medically accepted pediatric indications for

the 11 SGAs, we obtained copies of the compendium publications that would

have been the most recent available during our review period.35 Specifically, we

obtained the 2011 version of the American Hospital Formulary Service Drug

Information and the June 2011 DrugPoints.36

Collection of Medical Records

We developed a letter for our medical review contractor’s use in requesting

medical records from the prescribing providers identified on the sampled claims.

____________________________________________________________35 FDA-approved uses are included in the definition of medically accepted indications andtherefore are listed among the medically accepted indications in the compendia.36 The Act lists three compendia: (1) American Hospital Formulary Service Drug Information,(2) the United States Pharmacopeia—Drug Information (or its successor publications), and(3) DrugDEX Information System. However, we obtained copies of only the first and third ofthese. The second compendium listed in the Act, the United States Pharmacopeia—Drug Information, is no longer published. Its successor publication is DrugPoints, which containsinformation derived from DrugDEX —the third compendium.


http:///reader/full/period.35http:///reader/full/period.35http:///reader/full/period.35


26/40

For each of these providers, the contractor sent a letter requesting the medical

records associated with the sampled claim(s). If a provider did not respond to this

initial letter, the contractor sent up to two followup letters at predetermined time

intervals. For providers that did not respond to the initial letter or two followup

letters, the contractor followed up by telephone. We considered any provider that

had not submitted the requested records after being contacted at least five times(three letters and two telephone calls) to be a nonresponder.

Response Rates

Table C-2 shows how the response rate for the 485 claims for which we received

medical records were distributed among the 4 strata. Response rates varied by

stratum.

Table C-2: Response Rates by Stratum

Stratum Response Rate

Claims for children who were in foster care on the service date of the claim 62%

Claims for children who were not in foster care on the service date of the claimand for which the paid amount was less than or equal to $187

80%

Claims for children who were not in foster care on the service date of the claimand for which the paid amount was greater than $187 but did not exceed $557

73%

Claims for children who were not in foster care on the service date of the claimand for which the paid amount was greater than $557

69%

Source: OIG analysis of sampled claims and medical records received, 2014.

Nonresponse Analysis

Our findings are based on estimates made solely from medical records collected

for the associated sampled claims. We did not statistically adjust our results onthe basis of any assumptions about nonrespondent claims.

We performed a nonresponse analysis to determine the representation of

subpopulations in respondent claims and to determine the extent of the potential

nonresponse bias. We compared whether the respondents and nonrespondents

differed with regard to two variables that are available for the entire population

and that may be associated with response: (1) foster care enrollment and (2) State

of residence. We looked to see whether the response rates were significantly

different at the 95-percent confidence level. We estimated the proportion of

respondent and nonrespondent claims within these variables and performed

Rao-Scott chi-square tests of significance.We found significant differences in response rates for these two variables.

Records for children in foster care were less likely to have been submitted

(62 percent) compared with those for children not enrolled in foster care

(77 percent). Further, records for children residing in California were less likely

to have been submitted (65 percent) compared to those for children residing in the



27/40

other four States (79 percent). Therefore, claims for children in foster care and

claims from California are underrepresented among the respondent claims.

Potential Effects of Nonresponse on Estimates. To explore the effects that the

nonrespondent claims could have had on our results, we imputed and estimated

the following:

the percentage of claims that would have had at least 1 quality-of-care

concern if none of the 202 nonrespondent claims had had any

quality-of-care concerns and

the percentage of claims that would have been for drugs prescribed for

medically accepted uses if all of the 202 nonrespondent claims had been

for drugs prescribed for medically accepted uses.

If none of the 202 nonrespondent claims had had any quality-of-care concerns, the

percentage of all claims with at least 1 quality-of-care concern would still have

been at least 50 percent. If all of the 202 nonrespondent claims had been

prescribed for medically accepted uses, the percentage of all claims for drugs

prescribed for medically accepted uses would have been approximately

31 percent. See Table C-3 for 95-percent confidence intervals for these new

estimates based on the assumptions of errors in the nonrespondent claims.

Table C-3: Potential Effects of Nonresponse on Main Findings

Estimate DescriptionSample

SizePoint

Estimate95-Percent Confidence

Interval

Percentage of claims that would have had at least1 quality-of-care issue if none of the 202 claims forwhich we did not receive responses had had any

quality-of-care concerns

687 50.2% 44.7%–55.6%

Percentage of claims that would have been for drugsprescribed for medically accepted uses if all of the202 claims for which we did not receive responses hadbeen for drugs prescribed for medically accepted uses

687 31.5% 26.7%–36.7%

Source: OIG analysis of sampled claims and medical records received, 2014.

Review of Medical Records

Pretest of Medical Review Instrument . To test and refine the medical review

instrument, we pretested it with six randomly sampled claims from the MMIS

claims data from the selected States. For the pretest, we requested medical

records from the prescribing providers for the six sampled claims and asked the

medical reviewers to complete the medical review instrument. We then used the

reviewers’ feedback to improve the medical review instrument.

Determination of Medically Accepted Indications and Conditions Described in

the FDA Boxed Warning. Using (1) the diagnosis and behavior information that

the reviewers compiled from the medical records, (2) the age of each child with a



28/40

sampled claim, and (3) the compendia information, we determined whether each

claimed SGA was prescribed:

for medically accepted indications and/or

in the presence of conditions specified in the FDA boxed warning.

Development of Medical Record Review Criteria. Appendix B describes the

documents we reviewed to determine medical record review criteria for whether

children’s treatment with SGAs posed one or more quality-of-care concerns. The

criteria we identified include:

Dosage —All five documents support the idea that dosage needs to be

appropriate. Specifically, they say that children’s treatment should start

with the lowest effective dose; the dose should be adjusted and targeted;

and the dosage should not exceed any recommended dosage guidelines,

such as those for the FDA-approved uses.

Duration of use —Three of the five documents support the idea thattreatment with SGAs should be of appropriate duration. Specifically, they

say that a dose reduction should be planned after several months of

treatment, that the treatment plan should include a plan for discontinuing

the drug, and that medication trials should be of adequate duration to

assess the effects of the drug.

Indications for use —Four of the five documents support the idea that the

prescribed drug should be consistent with the child’s diagnosis. The

documents also say that there should be thorough and complete

evaluations of all aspects of a child’s condition and situation—including a

complete medical history and a psychiatric evaluation—and that the targetsymptoms and diagnosis for each drug that a child is prescribed should be

clearly documented. In most instances, psychosocial interventions should

be tried before starting treatment with drugs.

Monitoring —Four of the five documents support the idea that a

monitoring plan should be developed before the child starts taking an

SGA. The documents say that ongoing monitoring requires assessment of

the child’s clinical response to the drug; regular observation of

physiological changes, such as in height, weight, vital signs, or changes in

the results of bloodwork; and observation of any side effects experienced

by the patient, such as involuntary movements, insomnia, and other

physical and mental effects. The documents also say that that lack of

clinical drug trials and evidence-based data for pediatric uses of SGAs

reinforces the need for close monitoring of children taking these drugs.

Polypharmacy —All five documents support the idea that treatment with

a single drug should be tried before treatment with multiple drugs. They



29/40

say that there needs to be a clearly documented rationale regarding the

need for each drug when a child is treated with multiple drugs.

Side effects —All five documents support the idea that the use of SGAs

can have side effects. Some sources outlined specific measures to monitor

side effects, such as taking baseline and ongoing measures for height,

weight, blood pressure, and body mass index, as well as measuring

baseline and ongoing blood glucose and lipid levels.

Patient age —Four of the five documents support the idea that the

treatment of young children with SGAs should be rare and carefully

managed. Two of these four documents defined “young children” as those

under 4 years of age, another defined them as those under 6 years of age,

and the remaining document did not define an age.

To ensure a consistent interpretation of the seven criteria, we provided the

contracted psychiatrists with the same sources of guidance and information that

we reviewed to develop the criteria. The psychiatrists reviewed the medicalrecords for each sampled claim for an SGA and determined whether there were

quality-of-care concerns about the child’s treatment with the SGA with regard to

the seven criteria. Using their clinical experience and judgment, the psychiatrists

described any concerns they identified about the children’s treatment with regard

to any of the seven criteria.

Determination of Violations of State Coverage Policies

New York is the only State included in this evaluation that had a coverage policy

based on the compendia. New York’s policy limits Medicaid coverage to drugs

prescribed for medically accepted indications. To determine the extent ofcompliance with this policy, we identified all New York claims in the population

data that were for the five SGAs for which there were no pediatric medically

accepted indications.



30/40

APPENDIX D

Descriptive Data on Population and Sampled Claims

Tables D-1 through D-3 present descriptive information on the population of

claims, including the number of claims per State, each State’s foster-care

enrollment, and the SGAs prescribed. Table D-4 presents information on thetypes and specialties of the prescribing providers in the sample of claims. Table

D-5 presents estimates of the percentages and total numbers of claims in the five

States for which reviewers identified quality-of-care concerns.

Table D-1: Number of Claims and Children in Populationby State of Residence

State Number of Children Number of Claims forWith Claims for SGAs SGAs in Population

California 23,344 104,598

Florida 7,016 26,990

Illinois 15,704 74,329

New York 15,919 67,429

Texas 22,671 77,187

Total : 84,654 350,533

Source: OIG analysis of population claims data from five State MSIS systems, 2014.

Table D-2: Number of Children Enrolled in Foster Care With Claims in Population by State ofResidence

State

California

Florida

Illinois

New York

Texas

Number of ChildrenWith Claims While Not

in Foster Care

18,856

6,673

13,960

13,463

22,274

Number of ChildrenWith Claims While in

Foster Care

4,110

281

1,635

2,456

192

Number of Children With ClaimsBoth While i n Foster Care and

While Not in Foster Care

378

62

109

0

205

Total: 75,226 (89%) 8,674 (10%) 754 (1%)

Source: OIG analysis of population claims data from five State MSIS systems, 2014.



31/40

The table below provides information regarding the total number and percentage of claims in the population for each of the 11 SGAs. ThreeSGAs—risperidone, quetiapine fumarate, and aripiprazole—representednearly all claims in the population. This high representation is notunexpected. These three SGAs were among the first SGAs to be approved

by FDA. Additionally, these drugs are three of the five SGAs that have been approved by FDA for specific pediatric indications. Furthermore, theFDA-approved indications for two of these drugs—quetiapine fumarateand aripiprazole—have been expanded to include the treatment ofdepression associated with bipolar disorder and major depressive disorder.These factors likely contributed to the wider use of these three drugs overother SGAs.

Table D-3: Claims in the Population by SGA

DrugTotal

ClaimsPercentage of

Claims

risperidone* 164,636 47.0%

quetiapine fumarate* 91,830 26.2%

aripiprazole 49,360 14.1%

ziprasidone 21,924 6.3%

olanzapine 17,832 5.1%

paliperidone 1,650 0.5%

clozapine 1,482 0.4%

asenapine 1,296 0.4%

olanzapine/fluoxetine 330 0.1%

iloperidone 179 0.1%

lurasidone 14 0.0%

Total: 350,533 100.0%

*Includes an additional formulation, such as extended release, of thedrug.

Source: OIG analysis of population claims data from five State MSISsystems, 2014.



32/40

Table D-4: Prescrib ing Provider Type for Sampled Claims

Provider Type/SpecialtyNumber in Sample

Percentage of

Sample

Psychiatrist

Nonpsychiatrist MD

Non-MD*

Unknown

Total:

494

129

43

21

687

71.9%

18.8%

6.3%

3.0%

100.0%

*Examples of non-MD prescribing provider types included nurse practitioners,physician assistants, prescribing psychologists, and social workers.

Source: OIG analysis of sampled-claims data from five State MSIS systems andprovider type and specialty information from the National Provider Identifier Registry,2014.

Table D-5: Estimated Percentages of and Total Claims for SGAs withQuality-of-Care Concerns for the Five States Reviewed

Medical Record Review CriteriaPercentages of ClaimsWith Concern (n=485)*

Number of ClaimsWith Concern

Monitoring

Indications for use

Polypharmacy

Duration of treatment

Dosage

Patient age

Side effects

52.8%

40.6%

37.3%

34.1%

22.5%

16.7%

6.8%

138,679

106,674

97,986

89,542

59,123

43,924

17,773

*Percentages do not total 100% because medical reviewers identified multiple quality-of-careconcerns for many claims.

Source: OIG analysis of Medicaid claims data from five States and medical record reviews, 2014.



33/40

APPENDIX E

Statist ical Estimates and Confidence Intervals

Table E-1: Sample Sizes, Point Est imates, and Confidence Intervals

Estimate DescriptionSample

Size

Point

Estimate

95-Percent Confidence

Interval

Quality-of-Care Concerns (Percentages of Claims)

Percentage of SGA claims with at least one quality-of-careconcern

485 67.0% 60.6%–72.9%

Percentage of SGA claims with multiple quality-of-careconcerns

485 49.0% 42.5%–55.5%

Percentage of SGA claims with quality-of-care concernsrelated to monitoring

485 52.8% 46.3%–59.3%

Percentage of SGA claims with quality-of-care concernsrelated to indications for use

485 40.6% 34.5%–47.1%

Percentage of SGA claims with quality-of-care concernsrelated to polypharmacy

485 37.3% 31.3%–43.8%

Percentage of SGA claims with quality-of-care concernsrelated to duration of treatment

485 34.1% 28.2%–41.0%

Percentage of SGA claims with quality-of-care concernsrelated to dosage

485 22.5% 18.0%–27.8%

Percentage of SGA claims with quality-of-care concernsrelated to patient age

485 16.7% 13.0%–21.3%

Percentage of SGA claims with quality-of-care concernsrelated to side effects

485 6.8% 4.2%–11.0%

Medically Accepted Indications and Boxed Warning

Percentage of claims for SGAs prescribed for medicallyaccepted indications

485 8.5% 5.2%–13.6%

Percentage of claims for SGAs not prescribed for medicallyaccepted indications

485 91.5% 86.4%–94.8%

Percentage of claims for SGAs prescribed in the presenceof conditions listed in the FDA boxed warning

485 37.4% 32.3%–43.0%

Percentage of claims for SGAs that (1) were not prescribedfor medically accepted uses and (2) were prescribed to

treat bipolar disorder

459 20.1% 15.4%–25.9%

Percentage of claims for SGAs that (1) were not prescribedfor medically accepted uses and (2) were prescribed totreat mood disorders

459 13.0% 9.2%–17.9%

Percentage of claims for SGAs that (1) were not prescribedfor medically accepted uses and (2) were prescribed totreat autism spectrum disorders

459 8.4% 5.6%–12.5%

continued on the next page



34/40

Table E-1: Sample Sizes, Point Estimates, and Confidence Intervals (Continued)

Quality-of-Care Concerns (Numbers of Claims)

Number of SGA claims with quality-of-care concernsrelated to monitoring

485 138,679 119,568–157,790

Number of SGA claims with quality-of-care concernsrelated to indications for use

485 106,674 88,875–124,474

Number of SGA claims with quality-of-care concernsrelated to polypharmacy

485 97,986 80,469–115,502

Number of SGA claims with quality-of-care concernsrelated to duration of treatment

485 89,542 72,442–106,642

Number of SGA claims with quality-of-care concernsrelated to dosage

485 59,123 46,057–72,189

Number of SGA claims with quality-of-care concernsrelated to patient age

485 43,924 32,944–54,905

Number of SGA claims with quality-of-care concernsrelated to side effects

485 17,773 9,111–26,434

ource: OIG analysis of claims data and medical records, 2014.



35/40

PPENDIXF

gency Comments

. 1 - P ' ' ~

( ~ J _

DEPAR1MENT

OF

HEALTH

&

HUM N

SERVICES

Centers for Medicare

MediCaid

Services

.......

dministrator

Washington DC 20201

JAN 1 6 2015

DATE:

TO: Daniel R. Levinson

lnspectpr General

FROM: Marilyn a ~ e n n e r

Administrator

SUBJECT: Second-Generation Antipsychotic Drug Use Among Medicaid-Enrolled

Children: Quality

o

Care Concerns OEI-07-12-00320

The Centers for Medicare & Medicaid Services (CMS) appreciates the opportunity to review and

comment on the above subject draft report. This study provides CMS with important new

information to further our efforts to improve the quality of care provided to children with mental,

developmental, and/or behavioral health problems. CMS is committed to working with states,

other federal partners, and providers to address the significant quality

of

care issues identified in

the report.

As the draft report notes, during the period

of

January through June 30, 2011, the OIG examined

medical records from a stratified random sample

of

687 Medicaid fee-for-service paid claims to

determine the extent to which second-generation antipsychotics (SGAs) prescribed to children

enrolled in Medicaid in five selected states

1

: 1) presented at least one quality

of

care issue; and

2) were prescribed for medically accepted indications and/or prescribed despite the Food and

Drug Administration (FDA) boxed warning.

The

OIG found that the medical records for

67

percent of claims for SGAs prescribed to children

presented at least one quality

of

care issue in the five states reviewed. Lack

of

monitoring was

the most commonly identified quality

of

care issue; 53 percent

of

claims for SGAs prescribed to

children lacked documentation

of

monitoring in the five States reviewed. Furthem1ore, only 8

percent

of

SGAs were prescribed for medically accepted pediatric uses in the five states

reviewed; over one-third of SGAs were prescribed for conditions described in the FDA black

boxwaming.

CMS has been working over the past two years with its many partners, including the Medicaid

Medical Directors Network (MMDN), as well as other federal agencies such as the

Administration for Children and Families (ACF) and the Substance Abuse and Mental Health

Services Administration (SNvfi-ISA), to strengthen oversight and monitoring of psychotropic

medications use an1ong children. Examples of these efforts include:

1

California

Florida lllinois,

New

York and Texas

SGA Use Among Medicaid-Enrolled Children: Quality-of-Care Concerns (OEI-07-12-00320)

31

/S/


36/40



37/40


8/9/2019

gao second generation antipsychotic drug use on poor kids

Documents